Start your conversion optimization plan with your top-traffic pages, not everywhere at once. A small share of pages drives most of a site’s traffic and conversions, so a small improvement there produces the largest result, and those pages carry enough visitor volume for a test to reach a reliable answer instead of stalling on too few visitors.
Key takeaways
- Start with your top-traffic pages — they reach the most people and have enough volume for a reliable test.
- The 80/20 pattern holds for most sites: a small share of pages drives most of the traffic and conversions.
- Score each candidate page on traffic and performance gap to find the biggest, fastest wins, not a guess.
- Where is the first of three questions. What to test and why people convert come next.
- This isn’t the right starting point if you don’t have a landing-pages report to read yet, or if traffic itself, not conversion, is the actual constraint.
Is this your actual starting point?
Page-level prioritization only works once you have traffic to prioritize and a report that shows where it lands. Before ranking pages, check whether that’s genuinely where you are.
This isn’t your priority yet when:
- You don’t have analytics set up well enough to see a landing-pages report, in which case fixing measurement comes first, not picking a page.
- Total qualified traffic is still too low for any page to generate a meaningful sample, in which case the constraint is traffic, not conversion, and no page-selection method fixes that.
- You already know exactly which page is broken and why, in which case skip ahead to how to find where your funnel leaks instead of re-deriving what you already know.
If none of those apply, the rest of this guide is for you.
Where should you start optimizing?
Start with your top traffic pages, the ones visitors land on first from search, ads, email, and every other source, because a change there reaches the most people and has enough volume to test reliably. These pages carry the most opportunity for two reasons: more people see them, so a change reaches a bigger audience, and they get enough traffic for an A/B test to actually reach a reliable result instead of stalling on too few visitors.
This follows the 80/20 rule, also called the Pareto principle: most of the effect comes from a small share of the causes. Most websites fit this pattern. On an ecommerce site, the bulk of revenue comes from a small slice of the catalog, your best sellers. On a lead-generation site or a blog, the bulk of your traffic and conversions come from a small slice of your pages. That small slice is where your real conversion optimization opportunity lives. See the conversion rate optimization knowledge base for the deeper pattern library this prioritization method draws from.
How do you find those pages?
Use your analytics, specifically the landing-pages report, which shows the pages people arrive on first and tells you, without guessing, which pages do the heavy lifting. Almost every site finds its home page near the top of that report, alongside five to ten other pages pulling enough traffic to be worth optimizing.
Once you have that short list, narrow it with a few pointed questions. Which pages have the highest bounce rate? Which ones bring in plenty of traffic but convert poorly? Which ones drive the most revenue or the most leads? The pages where high traffic meets weak performance are your biggest, fastest wins.
Alt text: Four-quadrant matrix of traffic versus performance; the high-traffic, weak-performance quadrant is labeled “Start here.”
Figure 1 is the same logic as the questions above, made visual: a page only earns a top spot on your list when it sits in that upper-right quadrant, not just because it has traffic or because its performance looks weak in isolation.
Score your candidate pages
Score each shortlisted page 0 to 2 on three factors, then let the total tell you the order to work them in. This turns “which page feels most urgent” into a comparison you can defend to anyone who asks why you started where you did.
- Traffic volume. 0 = trickle, not enough for a reliable test. 1 = moderate, workable with a longer test window. 2 = strong, enough for a fast, reliable result.
- Performance gap. 0 = already converting well for its type. 1 = below average, but not alarming. 2 = high bounce, low progression, or a conversion rate well under similar pages.
- Business value. 0 = minor outcome if fixed. 1 = useful outcome. 2 = major revenue or lead impact if fixed.
A total of 5 or 6 goes to the top of the list. A 3 or 4 is worth a second look once the top scorers are underway. A 0 to 2 stays on the list but isn’t where you start.
What if a page doesn’t have enough traffic?
Be honest about volume before you commit: an A/B test needs a meaningful number of visitors to tell a real result from random noise, so low-traffic pages are not where a plan should start. Run one on a page with a trickle of traffic and you’ll wait a long time for an answer you still can’t trust. Lower-traffic pages aren’t hopeless, but they call for different methods, evidence gathered through user testing and careful, observed changes rather than a formal experiment.
What comes after “where”?
Finding where to start is the first of three questions: next comes what to test on those pages, then why people convert or don’t. Once you know which pages to focus on, the next question is what to test on them, which elements actually shape the decision a visitor makes. After that comes why: why people convert or don’t, why they hand over an email or walk away. The “where” gives you a focused list grounded in real data instead of a hunch.
Frequently Asked Questions
How many pages should I start with?
Most sites do best focusing on five to ten pages: the home page plus whichever other landing pages pull real traffic in your analytics report. Trying to work every page at once spreads effort too thin to move the number that matters; a short, ranked list keeps the plan focused.
What if my homepage isn’t actually my top traffic page?
That’s common and worth knowing early. Some sites get most of their volume through a specific landing page, a blog post, or a product page rather than the homepage. Let the landing-pages report decide, not an assumption about which page “should” matter most.
Do I need a special tool to find these pages?
No. The landing-pages report inside whatever analytics platform you already use (Google Analytics or similar) is enough to build this list. The skill here is asking the right questions of data you already have, not buying a new tool.
What to do next
- Pull your landing-pages report and shortlist the five to ten pages carrying real traffic.
- Score each shortlisted page 0 to 2 on traffic volume, performance gap, and business value; start with anything totaling 5 or 6.
- Don’t start on a low-traffic page even if it looks broken. A test there won’t reach a reliable answer.
- Once you have your starting page, move to what to test on it.
Continue based on what you need next
- You know your starting pages and want to know what to change on them: How to Find What to Improve on Your Website covers the second question, what to test.
- You’ve already fixed the obvious things and conversion still hasn’t moved: Why Your Conversion Rates Aren’t Higher covers the third question, why people convert or don’t.
- You want the full process this three-part sequence sits inside: CRO for Businesses With Traffic: A Plain-Language Guide is the cornerstone guide this page’s diagnostic is drawn from.
- You suspect the real problem is a specific step, not a whole page: How to Find Where Your Funnel Leaks goes deeper on step-by-step drop-off than the page-level scoring above.
Alex’s Perspective
The pattern that shows up across 100s of landing pages: teams almost always guess right about which page needs help, and almost always guess wrong about which one to fix first. The urgent-looking page and the highest-leverage page are rarely the same page, and a quick scoring pass catches that mismatch before a team spends a quarter optimizing the wrong one.
Where this fits
This is primarily a Prioritize decision: scoring and ranking candidate pages is how you decide which one deserves the next test, before you spend effort on any of them. Get the sequencing right here and the pages you actually work on are the ones most likely to move the number that matters.
Written by Alex Harris, who runs AlexDesigns’ conversion optimization work directly, page by page, rather than handing it to a team, which is why this guide describes the actual scoring method used to pick a starting page, not a general framework borrowed from somewhere else. Last reviewed 2026-07-18.
If you want a second set of eyes on which pages would repay the effort first, that’s exactly what a conversion review is for. Book a consultation and we’ll take a look.